def test_field_expansion(self):
print("Test deid.dicom.fields expand_field_expression")
from deid.dicom.fields import expand_field_expression
dicom = get_dicom(self.dataset)
contenders = dicom.dir()
print("Testing that field expansion works for basic tags")
expand_field_expression(dicom=dicom,
field="endswith:Time",
contenders=contenders)
print("Testing that field expansion works including private tags")
contenders += [e.tag for e in get_private(dicom)]
expand_field_expression(dicom=dicom,
field="endswith:Time",
contenders=contenders)
print("Testing that we can also search private tags based on numbers.")
fields = expand_field_expression(dicom=dicom,
field="contains:0019",
contenders=contenders)
# We should have a tag object in the list now!
assert isinstance(fields[0], BaseTag)
print("Testing nested private tags")
dataset = get_dataset("animals") # includes nested private tags
dicom = get_dicom(dataset)
def remove_private(self):
"""Remove private tags from the loaded dicom"""
try:
self.dicom.remove_private_tags()
except:
bot.error(
"""Private tags for %s could not be completely removed, usually
this is due to invalid data type. Removing others."""
% self.dicom_name)
for ptag in get_private(self.dicom):
del self.dicom[ptag.tag]
def dicom_dir(dicom):
"""Given a dicom file, return all fields (including private) if they
are not removed. With private this might look like:
...
'WindowCenterWidthExplanation',
'WindowWidth',
(0011, 0003),
(0019, 0010),
and both can be used as indices into the dicom (dicom.get(x))
"""
# This becomes a list of strings and tags to be used as keys
return dicom.dir() + [t.tag for t in get_private(dicom)]
def replace_identifiers(
dicom_files,
ids=None,
deid=None,
save=True,
overwrite=False,
output_folder=None,
force=True,
config=None,
strip_sequences=True,
remove_private=True,
):
"""replace identifiers using pydicom, can be slow when writing
and saving new files. If you want to replace sequences, they need
to be extracted with get_identifiers and expand_sequences to True.
"""
dicom_files, recipe, config = _prepare_replace_config(dicom_files,
deid=deid,
config=config)
# ids (a lookup) is not required
ids = ids or {}
# Parse through dicom files, update headers, and save
updated_files = []
for _, dicom_file in enumerate(dicom_files):
if isinstance(dicom_file, Dataset):
dicom = dicom_file
dicom_file = dicom.filename
else:
dicom = read_file(dicom_file, force=force)
dicom_name = os.path.basename(dicom_file)
# Remove sequences first, maintained in DataStore
if strip_sequences is True:
dicom = remove_sequences(dicom)
# Remove private tags at the onset, if requested
if remove_private:
try:
dicom.remove_private_tags()
except:
bot.error(
"""Private tags for %s could not be completely removed, usually
this is due to invalid data type. Removing others."""
% dicom_name)
private_tags = get_private(dicom)
for ptag in private_tags:
del dicom[ptag.tag]
continue
# Include private tags (if not removed) plus dicom.dir
fields = dicom_dir(dicom)
if recipe.deid is not None:
if dicom_file not in ids:
ids[dicom_file] = {}
# Prepare additional lists of values and fields (updates item)
if recipe.has_values_lists():
for group, actions in recipe.get_values_lists().items():
ids[dicom_file][group] = extract_values_list(
dicom=dicom, actions=actions)
if recipe.has_fields_lists():
for group, actions in recipe.get_fields_lists().items():
ids[dicom_file][group] = extract_fields_list(
dicom=dicom, actions=actions)
for action in recipe.get_actions():
dicom = perform_action(dicom=dicom,
item=ids[dicom_file],
action=action)
# Next perform actions in default config, only if not done
for action in config["put"]["actions"]:
if action["field"] in fields:
dicom = perform_action(dicom=dicom, action=action)
# Assemble a new dataset, again accounting for private tags
ds = Dataset()
for field in dicom_dir(dicom):
try:
# Most fields are strings
if isinstance(field, str):
ds.add(dicom.data_element(field))
# Remainder are tags
else:
ds.add(dicom.get(field))
except:
pass
# Copy original data attributes
attributes = [
"is_little_endian",
"is_implicit_VR",
"is_decompressed",
"read_encoding",
"read_implicit_vr",
"read_little_endian",
"_parent_encoding",
]
# We aren't including preamble, we will reset to be empty 128 bytes
ds.preamble = b"\0" * 128
for attribute in attributes:
if hasattr(dicom, attribute):
ds.__setattr__(attribute, dicom.__getattribute__(attribute))
# Original meta data # or default empty dataset
file_metas = getattr(dicom, "file_meta", Dataset())
# Media Storage SOP Instance UID can be identifying
if hasattr(file_metas, "MediaStorageSOPInstanceUID"):
file_metas.MediaStorageSOPInstanceUID = ""
# Save meta data
ds.file_meta = file_metas
# Save to file?
if save is True:
ds = save_dicom(
dicom=ds,
dicom_file=dicom_file,
output_folder=output_folder,
overwrite=overwrite,
)
updated_files.append(ds)
return updated_files
def replace_identifiers(dicom_files,
ids,
deid=None,
save=True,
overwrite=False,
output_folder=None,
force=True,
config=None,
strip_sequences=True,
remove_private=True):
'''replace identifiers using pydicom, can be slow when writing
and saving new files. If you want to replace sequences, they need
to be extracted with get_identifiers and expand_sequences to True.
'''
dicom_files, recipe, config = _prepare_replace_config(dicom_files,
deid=deid,
config=config)
# Parse through dicom files, update headers, and save
updated_files = []
for _, dicom_file in enumerate(dicom_files):
dicom = read_file(dicom_file, force=force)
dicom_name = os.path.basename(dicom_file)
fields = dicom.dir()
# Remove sequences first, maintained in DataStore
if strip_sequences is True:
dicom = remove_sequences(dicom)
if recipe.deid is not None:
if dicom_file in ids:
for action in deid.get_actions():
dicom = perform_action(dicom=dicom,
item=ids[dicom_file],
action=action)
else:
bot.warning("%s is not in identifiers." % dicom_name)
continue
# Next perform actions in default config, only if not done
for action in config['put']['actions']:
if action['field'] in fields:
dicom = perform_action(dicom=dicom, action=action)
if remove_private is True:
try:
dicom.remove_private_tags()
except:
bot.error(
'''Private tags for %s could not be completely removed, usually
this is due to invalid data type. Removing others.'''
% dicom_name)
private_tags = get_private(dicom)
for ptag in private_tags:
del dicom[ptag.tag]
continue
else:
bot.warning("Private tags were not removed!")
ds = Dataset()
for field in dicom.dir():
try:
ds.add(dicom.data_element(field))
except:
pass
# Copy original data attributes
attributes = [
'is_little_endian', 'is_implicit_VR', 'is_decompressed',
'read_encoding', 'read_implicit_vr', 'read_little_endian',
'_parent_encoding'
]
# We aren't including preamble, we will reset to be empty 128 bytes
ds.preamble = b"\0" * 128
for attribute in attributes:
if hasattr(dicom, attribute):
ds.__setattr__(attribute, dicom.__getattribute__(attribute))
# Original meta data # or default empty dataset
file_metas = getattr(dicom, 'file_meta', Dataset())
# Media Storage SOP Instance UID can be identifying
if hasattr(file_metas, 'MediaStorageSOPInstanceUID'):
file_metas.MediaStorageSOPInstanceUID = ''
# Save meta data
ds.file_meta = file_metas
# Save to file?
if save is True:
ds = save_dicom(dicom=ds,
dicom_file=dicom_file,
output_folder=output_folder,
overwrite=overwrite)
updated_files.append(ds)
return updated_files
def replace_identifiers(dicom_files,
ids,
deid=None,
save=True,
overwrite=False,
output_folder=None,
force=True,
config=None,
strip_sequences=True,
remove_private=True):
'''replace identifiers using pydicom, can be slow when writing
and saving new files'''
dicom_files, deid, config = _prepare_replace_config(dicom_files,
deid=deid,
config=config)
# Parse through dicom files, update headers, and save
updated_files = []
for d in range(len(dicom_files)):
dicom_file = dicom_files[d]
dicom = read_file(dicom_file, force=force)
idx = os.path.basename(dicom_file)
fields = dicom.dir()
# Remove sequences first, maintained in DataStore
if strip_sequences is True:
dicom = remove_sequences(dicom)
if deid is not None:
if idx in ids:
for action in deid['header']:
dicom = perform_action(dicom=dicom,
item=ids[idx],
action=action)
else:
bot.warning("%s is not in identifiers." % idx)
continue
# Next perform actions in default config, only if not done
for action in config['put']['actions']:
if action['field'] in fields:
dicom = perform_action(dicom=dicom, action=action)
if remove_private is True:
try:
dicom.remove_private_tags()
except:
bot.error(
'''Private tags for %s could not be completely removed, usually
this is due to invalid data type. Removing others.'''
% idx)
private_tags = get_private(dicom)
for ptag in private_tags:
del dicom[ptag.tag]
continue
else:
bot.warning("Private tags were not removed!")
ds = Dataset()
for field in dicom.dir():
try:
ds.add(dicom.data_element(field))
except:
pass
# Copy original data types
attributes = [
'is_little_endian', 'is_implicit_VR', 'preamble',
'_parent_encoding'
]
for attribute in attributes:
ds.__setattr__(attribute, dicom.__getattribute__(attribute))
# Retain required meta data
file_metas = getattr(dicom, 'file_meta', Dataset())
# Retain required meta data - not identifying
# file_metas.MediaStorageSOPClassUID
# file_metas.MediaStorageSOPInstanceUID
# file_metas.ImplementationVersionName
# file_metas.ImplementationClassUID
# File attributes for meta
attributes = [
'TransferSyntaxUID', 'FileMetaInformationGroupLength',
'FileMetaInformationVersion'
]
for attribute in attributes:
file_metas.add(dicom.file_meta.data_element(attribute))
# Preamble is required
ds.file_meta = file_metas
ds.preamble = vars(dicom)['preamble']
# Save to file?
if save is True:
ds = save_dicom(dicom=ds,
dicom_file=dicom_file,
output_folder=output_folder,
overwrite=overwrite)
updated_files.append(ds)
return updated_files